First belt of dead riverine jungle.
Continuous area of wind-eroded ground.
354 THROUGH THE LOP DESERT [Chap. X
brine-like that, in spite of the cold, it had nowhere yet frozen. The fact that such well-marked depressions seemed generally to stretch from north-east to south-west might suggest that they occupied patches of ground where wind-erosion had been particularly effective during some earlier and drier period. A narrow strip of ground which was crossed after a march of about eight miles, with living reeds and tamarisks, was proved by our surveys of 1914 and 1915 to extend several miles further to the west. In the same way a row of dead Toghraks between tamarisk-cones also dead, which we passed at about three miles further on, has been found to be in the continuation of an ancient river-bed traced in 1914 as coming from the north-west. In the light of our recent surveys we may, I think, now recognize in it the southernmost offshoot as yet traced of that ancient delta of the Kuruk-daryâ which brought water to the Lou-lan sites, and to which repeated reference will be made further on.
After passing through another large dry basin which seemed to have held water only for a short time during the latest period of exceptional inundation, we emerged at last in the evening on an area where wind-eroded banks and ` witnesses ', six or seven feet high, rose among low dunes. Bleached trunks of wild poplars and tamarisks of great size strewed the bare soil in abundance, unmistakably proving that a belt of luxuriant riverine jungle had once existed here. I received a clear impression of having left behind the terminal marsh region of the Târim for one where there was once running water and the possibility of tree growth. Shells of fresh-water snails lay scattered on the bare soil in plenty. About a mile and a half further on the dusk obliged us to pitch camp at the foot of high sand-cones covered with hoary tamarisk growth, some of which was still living. A well, dug in a hollow where the sand felt moist, yielded water at a depth of only five feet. It proved, as expected, utterly salt and undrinkable even for the camels. The vicinity, here demonstrated, of subsoil water suggested that water might have reached some of the previously passed saline depressions through percolation, and instances of this seemed well known to my Loplik guides.
SECTION III.-ACROSS AN ERODED ANCIENT DELTA
On the morning of December 15 I established, at Camp 121, the intended half-way depot by having all the bags of ice which had been brought on the thirty donkeys carefully stacked on the north side of the highest sand-cone. The donkeys, in charge of two extra men, were sent back by daybreak to Chainut-köl, whence after two days' rest they were to return with the labourers' reserve food supplies, as well as with fresh ice in the bags already emptied, and with some loads of reeds for the camels. A convoy of camels was to be sent back to this depot as soon as we had reached the ruins, for the purpose of fetching the remaining supplies. By careful timing I • endeavoured to assure that our divided transport columns should be saved unnecessary waits and the consequent risk of exhaustion. But my chief care had to be directed to the prompt location of the ruins.
This would necessarily depend on our steering correctly, and the difficulty was much increased by the marked change in the ground before us. From the very beginning of the march I realized that we had now definitely passed out of the level flat of recent lake basins, where tamarisk-cones and the like could serve as easily recognizable guiding points, and had entered a zone of a very different character. Throughout the day's march, which by a great effort was extended to close on sixteen miles, we crossed ground cut up into a bewildering succession of wind-eroded ridges and trenches, where it was not covered by low dunes. The soil was a very hard greyish clay, and the )(di-clangs which the wind had carved out of it ran regularly from east-north-east to west-south-west. The trenches dividing them always showed steep banks, the depths varying from about four to ten feet. The top of the terraces or plateaus left between these trenches was also carved by a network